1. Evaluating the impact of ocean gravity wave variability on Aquarius satellite measurements D. Vandemark, H. Feng Univ. of New Hampshire/EOS N. Reul, F. Ardhuin, B. Chapron IFREMER/Centre de Brest within Aq. Cal/Val team efforts D. Vandemark, H. Feng Univ. of New Hampshire/EOS N. Reul, F. Ardhuin, B. Chapron IFREMER/Centre de Brest within Aq. Cal/Val team efforts OSST Meeting 2012 2OSST 2012

2. Overview Goal Develop and refine an empirical satellite salinity correction for long wave impacts that augments the 1 st order roughness corrections made using NCEP winds, Aquarius scatterometer, or other ocean roughness information Approach Geolocate ancillary ocean wave data with each satellite data point Detect, characterize, and quantify long-wave impacts seen in the scatterometer, radiometer, and ultimately salinity Help implement a point-by-point correction using operational wave model data 3OSST 2012

3. 4

4. 5 The general geophysical problem To obtain accurate salinity we need to accurately remove the signal due to rough surface emission Specifically for Aquarius 1) A three beam radiometer at L band - Tb_ocean (2 polarizations, 3 incidence angles, look angle) = F ( variable sea surface waves, S(k) ) ~= rms slopes = 2) A three beam radar scatterometer at L-band (Bragg waves ~ 18 cm) NRCS_ocean (2 polarizations, 3 incidence angles, look angle) = F ( variable sea surface waves ) ~= S(k Bragg ) + tilting effects 3) The portion of the wave spectrum and EM interaction differs for the radar and radiometer and for each beam’s incidence angle – how well correlated are 1) and 2)? 4) Usual surrogate for sea surface wave information is wind – unlikely to provide sufficient precision to correct for the true ocean wave field.

5. OSST 20126 Global Wave model fields, Aquarius L2_wwav files, Day 240-271 2011 Aquarius *L2_wwav files are assembled daily at UNH for the Aq. cal/val team available at PO.DAAC

6. Aquarius Level 2 wave model collocation products: Processing status (March 2012) Near Real Time processing ( Latency: 1 day after Aq-L2 V1.1 data become available at GSFC) Aq-L2 V1.1 SCIdata (daily wget) L2V1.1_wwav* Ifremer/Previmer WW3 model product ( Partial dataset: 5 NRT variables, limited QC) Science level processing ( Latency: 50-60 days after Aq-L2 data become available) L2V1.2_wwav (2011)* L2V1.2.3_wwav (2012)* Ifremer/Previmer WW3 model product (Full wave model dataset: 14 variables with QC) Aq-L2 V1.2 EVSCI data (2011) Aq-L2 V1.2.3 EVSCI data (2011, 2012) * ftp access via PODAAC for Aq. Cal/Val team members

7. OSST 20128 Diagnosis using Significant Wave Height (SWH) Evaluate Aquarius L2 scatterometer and SSS data Expectations: 1)For fixed wind speed we’ll see long wave impacts in the scatterometer roughness and derived wind speeds 1)For fixed wind speed we’ll see long wave impacts in the radiometer-derived salinity

8. Aquarius Scatterometer sigma0 vs. significant wave height (17 weeks; Day 240-362) ; X-axis=NCEP wind OSST 2012 VV POL INNER BEAM θ = 28 deg. HH POL INNER BEAM θ = 28 deg. VV POL OUTER BEAM θ = 46 deg. HH POL OUTER BEAM θ = 46 deg. LARGEST IMPACT LEAST IMPACT 9

9. Aquarius Scatterometer Wind vs. significant wave height (17 weeks; Day 240-362) ; X-axis=NCEP wind OSST 2012 INNER BEAM θ = 28 deg. OUTER BEAM θ = 46 deg. LARGEST IMPACTWEAKER IMPACT Long-waves lead to wind speed error when seas are exceedingly high This is a good thing – indicates that the scatterometer is sensitive to longer waves that likely impact the radiometer 10

10. Getting sidetracked OSST 201211 All wind products show significant systematic differences – likely associated with ocean currents. wind waves, and atmospheric stability (SST) impacts. NCEP, ECMWF, SSM/I, Aquarius Scatterometer

11. Wind products and spatial differences Day 240-270 OSST 201212

12. Wind products and spatial differences Day 240-270 OSST 201213

13. Onto radiometer derived SSS OSST 201214 Some substantial improvement already L2 V1.2 -> 1.3

14. Aquarius Radiometer Salinity error vs. SWH (17 weeks; Day 240-362) ; X-axis=NCEP wind, V`1.2; DESC OSST 2012 Residual Salinity with respect to HYCOM Long-waves lead to salinity anomaly – low for low seas, high for high seas Residual Salinity taken with respect to HYCOM model SSS 15

15. Aquarius Radiometer Salinity error vs. SWH (17 weeks; Day 240-362) ; X-axis=NCEP wind; Ver 1.2.3; DESC; OUTER BEAM Aq Cal/Val, March 2012 Residual Salinity with respect to HYCOM Long-waves lead to salinity anomaly – low for low seas, high for high seas Residual Salinity taken with respect to HYCOM model SSS

16. Aquarius Scatterometer Wind vs. significant wave height (17 weeks; Day 240-362) ; X-axis=NCEP wind; V1.2 OSST 2012 OUTER BEAM θ = 46 deg. Desc Pass data, Galactic refl < 1 K Long-waves lead to wind speed error of order when seas are exceedingly high OUTER BEAM θ = 46 deg. 0.35 psu/ 1m sea state change 0.21 psu/ 1m sea state change 17 DESCENDING V1.2 Long-waves can lead to wind speed error of order when seas are exceedingly high

17. Aquarius Scatterometer Wind vs. significant wave height (17 weeks; Jan-Feb2012) ; X-axis=NCEP wind; Ver. 1.3 Aq Cal/Val, March 2012 OUTER BEAM θ = 46 deg. Desc Pass data, Galactic refl < 1 K OUTER BEAM θ = 46 deg. 0.1 psu/ 1m sea state change <0.1 psu/ 1m sea state change DESCENDING V1.3 Substantial improvements

18. Aquarius Scatterometer Wind vs. significant wave height (17 weeks; Jan-Feb2012) ; X-axis=NCEP wind; Ver. 1.3 Aq Cal/Val, March 2012 OUTER BEAM θ = 46 deg. ASC Pass data, Galactic refl < 1 K OUTER BEAM θ = 46 deg. >0.1 psu/ 1m sea state change ~0.1 psu/ 1m sea state change ASCENDING V1.3 But not done yet

19. Some caveats and future work Final Aquarius surface Tb_V and Tb_H are still being developed with issues such as reflected Galaxy and Faraday rotation correcting impacting these 2nd order ocean wave impact results (V1.4?) ascending vs. descending results differ time dependence highest wind/wave environment are in coldest waters SST-wave covariance Fully exploit and document the certain benefit of using the scatterometer for SSS inversion Further formal evaluation of the radiometer and radar data within a scattering/emission model + global wave field data What wind model to use as a reference? SSM/I != NCEP != ECMWF 20OSST 2012

20. Analysis of Satellite SSS signatures over Hurricane IGOR 11-24 Sep 2010 using SMOS Cat 4-5

21. Freshwater Plume=> warmer SST & shallow stable surface layer (barrier layer) 65% of TC crossing the Amazon Plume evolve into cat 5 Hurricanes Ffield (J. Clim 2007) Amazon & Orinoco plumes =>Strong positive SST anomalies (~1°C) Impact of plume SST on storm intensity addressed using WRF simulation Vizy and Cook, (JGR 2010) Can we use SMOS retrieved SSS to study ocean-atmosphere interactions in TCs ? Nico Reul with Chapron, Tenerelli, Vandemark, Vialard,...

22. Analysis of SMOS data signature over Hurricane IGOR 11-24 Sep 2010 Cat 4-5

23. Location of the Plume (SSS<35.5) the last 10 days before Igor passage (5->13 Sep) 09/13/2010 09/15/2010 09/17/2010

24. are cross-track locations from the eye at t o Strong resalinisation of the SSS after Igor passage, over the path of the storm Right-hand quadrant: ΔSSS~>+1-1.5 pss Cool Wake induced by Igor: SST= GHRSST OSTIA (MetOffice) Salty Wake induced by Igor: SSS=SSS SMOS (CATDS/ifremer)

25. SSS averaged 5-11/09 (3-5 days before Igor ) SSS averaged 19-24/09 (3-5 days after Igor ) Apparent Erosion of the freshwater surface layer on the right-hand side quadrants of the storm ARGO profiler #4900819 7/09 ARGO profiler #4900819 17/09

26. ΔSSS smos =+1.4 ΔSSS argo =+1.3 Enhancement of Sea Surface Salinity as seen by ARGO after IGOR passage: ΔSSS argo ~+1.3 Excellent consistency with SMOS observation trend: ΔSSS smos ~+1.4. SMOS SSS is however systematically ~0.5 pss fresher than ARGO observations at 5 m depth Maximum surface Wind encountered at the argo float location is ~33 m/s (GFDL model) Maximum significant wave height up to 11 m (Wave Watch III, NAH)

27. Original Plume surface freshwater layer ~20 m thick After the passage of Igor: =>Drop of sea surface temperature Δ T~2.8°C =>Enhancement of Sea Surface Salinity by ΔS=1.4 Excellent consistence with SMOS obs~1.5 Increase in sea surface density ~2 kg/m 3 Mixing of the Warm freshwater surface layer by Hurricane Igor (ARGO#4900819) 6 days Before Igor 1 day after Igor 9 days After Igor Deepening of the Mixed surface layer From 20 m down to ~90 m depth Salinity Argo #4900819 Temperature Argo #4900819 Density Argo #4900819

28. ARGO profiler #6900590 4/09 SSS averaged 5-11/09 (3-5 days before Igor ) SSS averaged 19-24/09 (3-5 days after Igor ) ARGO profiler #6900590 24/09

29. ΔSSS argo =+0.3 ΔSSS smos =+0.4 Surface wind speed encountered >15 m/s up to 48 m/s from 09/15->09/17 Significant wave height >3 m up to 9 m from 09/15->09/17 SMOS SSS in general saltier than Argo SSS by ~0.05-0.1 pss However very consistent SSS temporal trend from before to after Igor between ARGO float & SMOS surface data: =>both are showing a +0.3-0.4 pss increase following the surface mixing induced by IGOR on its left hand side quadrant.

30. 9 days Before Igor 1 day before Igor 6 days After Igor Cooling ~1°C More intense wave mixing On the RHS ? Mixing Damped By the thicker Plume Barrier layer On the LHS ?

31. EARLY CONCLUSIONS SMOS SSS data able to produce before and after snapshots of plume location associated with TC passage Satellite SSS data yielding accurate SSS perturbation due to TC as compared to two ARGO floats – 0.5 to 1.5 SSS INCREASE New look at plume-TC interaction with SSS + SST perhaps allowing enhanced diagnosis

32. Figure 4: Surface wakes of Hurricane Igor. Post minus Pre-hurricane (a) Sea Surface Temperature (ΔSST ) (b) Sea surface Salinity (ΔSSS), (c) Sea Surface Density (Δσ o ) and (d) Sea Surface CDOM absorption coefficient.The thick and thin curves are showing the hurricane eye track and the locii of maximum winds, respectively. The dotted lines is showing the pre-hurricane plume extent. ΔSST, ΔSSS, Δσ o wakes were only evaluated at spatial locations around the eye track for which the wind exceeded 34 knots during the passing of the hurricane. Surface wakes of Igor Six days of data centered on t o –(+) 4 days have been averaged to construct the pre (post)- cyclonic quantities. Here a cdom = a d + ag ag: CDOM (dissolved matter) ad: non living particulate organic material, bacteria, inorganic material and bubbles

33. Thanks! 34OSST 2012 This work is supported by NASA’s Ocean Surface Salinity Science Team – Grant NNX09AU69G